Low-Cost, Continuous Motion Imaging, Computationally Augmented Whole Slide Imager for Digital Pathology

Prateek Katare, Navchetan Awasthi, Aravind Venukumar, Sai Siva Gorthi (Corresponding author)

Onderzoeksoutput: Bijdrage aan tijdschriftTijdschriftartikelAcademicpeer review

1 Citaat (Scopus)


As digital pathology advances, systems for acquiring Whole Slide Images (WSI) need to become affordable and effective. Current research is majorly reported around the low-throughput stop-and-focus imaging systems that are too slow for clinical use. This work reports a system that acquires well focused images at a high throughput. It reduces the imaging time by an order of magnitude via the carefully engineered implementation of a global shutter camera which allows continuous motion imaging. The second challenge of maintaining good focus comes from three sources: the imperfections induced by low-cost optomechanics, smear surface undulations, and irregularity of the slide surface itself. These are minimized via a tiled-scan method supported by focal mapping, which predicts and achieves best focus on-the-fly. Finally, any remnant defocus is compensated for by a post-acquisition image enhancement method which sharpens the image without disturbing the essential features and color tones. Thus, this system delivers a 15 × 15 mm2 scan in about 3 minutes at a resolution of 0.78 μ m with a 40x microscopy objective. The bill of material cost is about US 1300 and hence it would be beneficial for telepathology in resource-constrained scenarios. Also, this device can serve as an enabler of Artificial Intelligence/Machine Learning algorithms to provide fully automated diagnosis. The developed codes for post-acquisition image enhancement are available at https://github.com/navchetan-awasthi/Microscopy.

Originele taal-2Engels
Aantal pagina's7
TijdschriftIEEE Journal of Selected Topics in Quantum Electronics
Nummer van het tijdschrift4
StatusGepubliceerd - 1 jul 2021


Duik in de onderzoeksthema's van 'Low-Cost, Continuous Motion Imaging, Computationally Augmented Whole Slide Imager for Digital Pathology'. Samen vormen ze een unieke vingerafdruk.

Citeer dit